Data Scientist Salary
February 2, 2023
Philippe Lambot — GitHub Account
Thanks to ai-jobs.net for its full dataset of global AI/ML salaries, which is published in the public domain under the CC0 license. The CSV file has been downloaded on January 15, 2023, at 01:15 CEST.
Throughout this data analysis project, salary is — unless otherwise stated — total gross salary paid. All salaries are denominated in USD for the sake of comparability.
The graph above shows the evolution over the whole period 2020-2022 in the three continents that are well represented in the dataset under review; unfortunately, the dataset has not been provided with numerous positions from the two remaining continents, which are therefore not represented in the graph but whose countries will show in all comparison tables by country.
Let us dig up insights from the graph.
In the graph above, dragging the animation slider reveals that
- the sets of points grow year by year;
- the sets of points of the continent America are by far more populated than the other ones, at least in 2021 and 2022;
- as shown by the violin shapes, all 2022 median salaries are higher than the 2020 medians;
- every year, continent medians rank quantitatively in the order America-Europe-Asia with big gaps between them;
- hovering over the overall highest point — reported for 2020 — shows it is a salary of USD 450,000 earned by a Research Scientist with mid-level experience working full-time in a medium-size company in the United States with a remote ratio smaller then 20% and with the United States as primary country of residence;
- the overall lowest salary is in 2021, with an amount of USD 5,000 for a 3D Computer Vision Researcher at senior level working full-time for a large company in the United States with a remote ratio greater than 80% and with the United States as primary country of residence.
Hovering over the 2022 violin shapes shows median is around
- USD 142k in the continent America,
- USD 63k in Europe,
- USD 24k in Asia.
Hovering over these shapes also shows around 50% of salaries are
- between USD 110k and USD 180k in the continent America,
- between USD 47k and USD 85k in Europe,
- between USD 14k and 53k in Asia.
Hovering over the 2022 sets of points shows that
- at the top of the sets of points, most positions are at senior or executive level;
- the set of points of the continent America is mainly comprised of positions in companies located in the United States;
- above the upper fence in the continent America — namely USD 275,000 —, there is a subgroup of salaries between USD 297,500 and USD 405,000, at senior or executive level, in the United States;
- above the upper fence in Europe — namely USD 135,000—, there are four outliers, three of them in the United Kingdom and one in Germany;
- one of the outliers in Europe is a super outlier equivalent to USD 430,967 for a Data Analyst with mid-level experience, working full-time with a remote ratio lower than 20% in a medium-size company located in the United Kingdom, and having the United Kingdom as her/his primary country of residence.
The evolution over time of median salaries by continent shows separately in the next graph.
When comparing continents, gaps appear in terms of median salaries and — to a lesser extent — in terms of trends over time.
In 2022, median salary was the equivalent of
Ratios between continent medians are
Trends partially diverge in that
The next graph shows the evolution over time of average salaries by continent. In this Data Analysis project, the word "average" refers to the concept of statistical mean.
The gaps between continents — which had been observed in terms of medians — are found back here in terms of averages.
In 2022, average salary is approximately the equivalent of
Nevertheless, there are some differences between the vision given by the medians and that given by the averages. Indeed, each concept reflects reality in a different way. Averages lie higher than medians, especially so in Asia and in the continent America for 2020:
Indeed — as the first graph shows —, the distributions are both scarce and elongated for Asia and the continent America in 2020, so that some high salaries can heavily swell averages without impacting medians.
For 2021, the gaps between median and average are substantially reduced for Asia and the continent America. Distributions are already denser in both continents, and — for Asia — the outlier of 2020 equivalent to USD 260,000 does not recur in 2021.
Indeed reminds us here of the respective merits of median and average.
In the interactive table above — which has been sorted by frequency — the job title of Data Engineer appears first, followed by Data Scientist, Data Analyst, and Machine Learning Engineer. These four roles will be studied in four separate sections — sections 3 to 6 —, after having each been amalgamated with similar job titles.
Hovering over the sets of points in the first graph had indicated some correlation between high salaries and senior or executive experience levels. In the dataset under review, which are the frequencies of the various experience levels — with their position counts — for 2022?
As shown in the table above, the most represented experience level is the senior one. That can solidly contribute the statistical representativeness of median and average estimates at that level.
Another feature in the dataset under review is the type of employment. Here are the frequencies of the different employment types.
The full-time employment type is absolutely predominant, its prevalence exceeding 99%. Only this type of employment will be analyzed in sections 3 to 6. The other three types will not be included because they are so scarcely represented in the dataset under review.
Last but not least, the breakdown by country appears in the next table.
As this has been seen when hovering over the sets of points in the first graph, the most represented country is by far the United States. The table above shows that the United States is followed, at a great distance, by the United Kingdom, Spain, and Canada. This ranking may differ in sections 3 to 6 depending on the job title group or on the job title.
In a snapshot about frequencies, the highest ones relate to
- full-time jobs,
- the United States,
- the senior level of experience,
- and the job titles of Data Scientist, Machine Learning Engineer, Data Engineer, and Data Analyst.
In this data analysis, salaries are annual total gross salaries paid — unless otherwise stated —, and they are denominated in USD for the purpose of comparability.
This section will deal with all job titles containing the expressions "Data Scientist" or "Data Science", which can be pinpointed in this table for 2022.
Machine Learning — which intersects with Data Science (DS) and Artificial Intelligence — will be dealt with separately in the next section.
The country comparison table above relates to all — full-time as already stated — Data Science job titles in 2022.
Average salaries are somewhat higher than median salaries, except of course when the two concepts merge because there are only one or two positions reported...
Sorting the table by median salary in descending or ascending order gives the highest or lowest salary medians by country.
The highest medians by country are, in descending order, those of
- the United States — with USD 150,000 for 371 positions reported,
- followed by Switzerland — with one salary equivalent to USD 120,402,
- Australia — with a median salary equivalent to USD 83,171 for three positions,
- the Netherlands — with a median equivalent to USD 81,426 for four positions, ...
The lowest medians by country are those of
- Singapore — with one salary equivalent to USD 8,000,
- Turkey — with a median equivalent to USD 17,500 for two positions,
- Hungary — with one salary equivalent to USD 17,684,
- India — with a median salary equivalent to USD 22,892 for four positions,
- Italy — with one salary equivalent to USD 24,165, ...
All Data Science job titles are dealt with in this section. Besides the job titles of Data Scientist and Data Science Manager, the other Data Science job titles are listed in this table for the United States in 2022.
Based on the dataset under review, the median salary of the group of Data Science job titles has evolved in the United States as shown in the graph below.
Median salary — represented by the central horizontal bar — is ascending from 2020 to 2022 — especially from 2021 to 2022. Median salary is USD 150,000 in 2022. The violin plots deliver other salary summary statistics.
With a view to extracting more insights about 2022 from the graph above, let us enlarge the part of the graph relating to 2022.
The Data Science violin figure above shows for 2022 that
Some summary statistics are transcribed in the next table for 2022. The table offers two subgroups of salary statistics: in a first subgroup, salary statistics are centralized around the median, with first and third quartiles, minimum and maximum; a second subgroup is comprised of the average and the number of positions.
Averages are sensitive to outliers. The average salary of Data Science jobs in the United States is USD 156,794 for 2022 — against USD 150,000 for the corresponding median. This average salary — which is based on annual total gross salaries paid, as already mentioned — seems rather in line with an estimate on Indeed, consulted on December 28, 2022: the average base salary of Data Scientists was estimated at USD 145,150.
These statistics have been established for the United States as a whole since the dataset under review is comprised of neither State nor city identification. But on the Internet, there are indications of disparities between cities. For instance, Indeed — consulted on December 28, 2022 — indicated, even just among the highest paying cities, a substantial difference in average base salary : USD 161,049 in Houston, TX, against USD 120,324 in Redmond, WA. Furthermore, when comparing States, the same study on Indeed shows diversity: in the highest-paying States, base salary average is 67% higher than in the lowest-paying States.
On Forbes, contributor Andrew DePietro rightly refers to the U.S. Bureau of Labor Statistics. Indeed, the U.S. Bureau of Labor Statistics delivers in-depth statistics about, among others, Data Scientist salaries. Their Data Tools allow to download customized tables. I have customized and downloaded four XLS files about Data Scientist salaries, from the May 2021 version:
In the lowest-paying State, the average salary is in a ratio of 1 to 2.1 with respect to the highest-paying State: USD 64,360 in West Virginia, against USD 133,900 in the State of Washington. Disparity increases when switching to MSA level: in South Central Kentucky nonmetropolitan area, the average salary of USD 53,390 is in a ratio of 1 to 2.9 with respect to the average of USD 157,110 in San Jose-Sunnyvale-Santa Clara, CA metropolitan area.
The U.S. Bureau of Labor Statistics mentions that it obtains annual average salaries by multiplying hourly average salaries by 2,080. I have done the same to convert hourly median salaries to annual median salaries. Calculated in that way, the annual median salary is USD 129,563 in the State of Washington, that is to say 110% higher than USD 61,714 in West Virginia. When comparing MSAs, the range is larger. In South Central Kentucky nonmetropolitan area, the median salary is in a ratio of 1 to 3.8 with respect to San Jose-Sunnyvale-Santa Clara, CA metropolitan area: USD 42,682, against USD 163,238.
Readers interested in additional information, for instance in job concentration by geographical unit, are invited to refer to the in-depth statistics and analysis by the U.S. Bureau of Labor Statistics.
After having somewhat quantified the link between company location and salary, let us turn to the breakdown by experience level.
As shown in the violin-shaped figures above, median salary follows an upward trend from junior to executive level. Hovering gives
So, median salary is in a ratio of 1 to 1.9 at junior level with respect to executive level. This ratio lends support to the idea of experience level being an impactful salary determinant.
To gain a broader perspective, it could be remembered, however, that the ratio is 1 to 3.8 when comparing MSAs (metropolitan statistical areas) inside the United States, not to mention higher ratios between countries, such as 1 to 4 between Spain and the United States, 1 to 6.6 between India and the United States, 1 to 8.6 between Turkey and the United States, ...
The senior level median of USD 155,000 seems compatible with estimates given on Built In — consulted on December 28, 2022: for Senior Data Scientists in the United States, Built In gives a median of USD 140,000 for base salary and an average of USD 21,243 for additional cash compensation.
In 2022, senior level is rather densely populated and is by far the most populated, which comforts the statistical reliability of the estimated median at senior level. In 2021, positions reported were scarce.
Now, let us drill down at a micro level by hovering over the sets of points.
At the top of the sets of points, let us notice that
At the bottom of the mid-level set of points, there are two salaries corresponding to primary countries of residence other than the United States: there is a salary of USD 31,795 and one of USD 48,000 — with respectively India and Russia as primary country of residence. This raises a question: would residence influence the salary amount? All cases of this kind will be analyzed here.
After data mining into the sets of points corresponding to experience levels, let us fine tune this analysis by combining job titles and experience levels. The job title of Data Scientist comes first, being the most frequent.
The median salary of Data Scientists increases from junior level to senior level, and then decreases somewhat based on two positions.
Salaries at senior level show a bimodal distribution. There is concentration at two quantitative niveaux, with two peaks:
Such bimodality does not show clearly in standard summary statistics such as median, first and third quartiles, 10th and 90th percentiles, etc.
Medians by experience level are transcribed in the next table, sorted by position count in descending order.
Senior level is clearly predominant — in both position number and median salary.
The next table shows the evolution over time of the median salary of Data Scientists.
Median salary makes an impressive jump in 2022, based on available observations.
Data Science Manager is the second job title — after Data Scientist — in terms of occurrence.
Here are some salary summary statistics about the job title of Data Science Manager by experience level in the United States in 2022.
Based on available observations, the table above shows that
Comparing level by level the median salaries of Data Science Managers and Data Scientists shows that Data Science Manager medians noticeably surpass Data Scientist medians — on the basis of available data. For instance, at senior level — which is more populated —, the median salary of Data Science Managers is USD 185,400, against USD 152,500 for Data Scientists.
After visualizing statistics about the job titles of Data Scientist and Data Science Manager, let us visualize all combinations of the other Data Science job titles with experience levels.
By default, the table above is sorted by position count in descending order and then by median salary. Data Science Consultant at senior level comes first with 6 positions.
Sorting rows by maximum salary in descending order gives
After breakdowns by company location, job title, and experience level, here is a breakdown by company size for all Data Science job titles taken together.
Median salary by company size ranges over two quantitative niveaux:
How densely are the three subsamples populated? In 2022, the small company subsample is scarcely populated – with four positions –, the large company subsample is somewhat larger, but most positions show in the medium-size company subsample. From a sample size viewpoint, the medium-size company median gets more statistical reliability than the small-size company median.
The data distribution of the medium-size companies is slightly bimodal, revealing some concentration at two quantitative niveaux. The data distribution of the large companies has a left shoulder, which corresponds to a lower subgroup of salaries, without any low salary being considered an outlier.
Let us now turn to another feature of job positions: remote ratio.
After dealing with company location, job title, experience level, and company size, let us turn to another feature, namely remote ratio. Will the salary breakdown by remote ratio be marked by sharp disparities?
Median salary evolves according to a caret-shaped profile.
Median salary lies almost at the same quantitative niveau whether in the salary subsample with remote ratios smaller than 20% or in the salary subsample with remote ratios greater than 80%. Hovering over the graph boxes shows medians of respectively USD 150,000 and USD 152,500.
The intermediate category has a somewhat higher median salary, namely USD 165,000.
But to what subsample sizes do these medians correspond?
The intermediate category is hardly populated in 2022. Almost all positions except for three of them are reported in the extreme categories, whose summary statistics offer – from a sample size viewpoint — much more statistical reliability.
In 2020, no position was reported with a remote ratio of less than 20%: let us not forget that the Covid-19 pandemic was spreading. In 2021, the biggest subgroup was the one with remote ratios greater than 80%, the other two subgroups having almost the same number of positions.
In the analysis above, hovering over low salaries has sometimes shown an employee's primary country of residence differing from the company location country. For the sake of clarity, analyzing the impact of such a difference has been postponed so that it did not bypass the analysis parts that were in progress.
Actually, over the whole period, there are only seven positions with the employee's primary country of residencea differing from the company location country — namely the United States. Six positions correspond to the job title of Data Scientist, one to the job title of Data Science Manager.
A picture being worth a thousand words, let us use graphs to situate these cases in broader and relevant subsamples.
They will be analyzed in the following order:
Let us get started with the two positions of Data Scientist at mid-level from 2022.
In 2022, there are two positions of Data Scientist at mid-level with a difference between the employee's primary country of residence and the company location country:
They are both at the bottom of the set of points of 2022. The two values just above are USD 78,000; the first quartile is USD 101,050; the median salary is USD 120,000.
Actually, the salary of USD 31,795 is closer to an average equivalent to USD 22,256 reported for two Data Scientist positions at mid-level in India, as shown in this table.
In 2021, there is a salary of USD 5,679 for a Data Scientist at mid-level, working in a small-size company, and having India as her/his primary country of residence. In order to situate this salary of USD 5,679 among other salaries, let us notice in the graph just above that
Actually — just as for the salary of USD 31,795 analyzed above — the salary of USD 5,679 is closer to an average salary equivalent to USD 22,256 and corresponding to two Data Scientist jobs at mid-level in India, as shown in this table.
In 2020, there is one salary of Data Scientist at mid-level: USD 45,760 in a small-size company, with the Philippines as primary country of residence. That salary is by far the lowest one in the set of points of 2020, as shown in the graph just above.
There are also two salaries of Data Scientist at senior level:
Let us get started with the salary of USD 100,000 in 2022 by situating it in a relevant subsample thanks to a reference graph.
The violin graph above helps us situate the USD 100,000 salary vis-à-vis American references from the dataset under review:
Let us now try situating the USD 100,000 salary with respect to Brazilian salaries:
In short, the USD 100,000 salary is in between the senior level median of USD 152,500 for the United States and estimates for Brazil.
The next graph situates the salary of USD 68,428 — for a Data Scientist at senior level in 2020 — alongside other positions of Data Scientist at senior level with the United States not only as company location but also as primary country of residence of the employee.
The salary of USD 68,428 is below
Last, there is a salary of USD 54,094 in 2021 for a Data Science Manager at senior level in a large company, with India as primary country of residence but with a remote ratio between 20% and 80%. The next graph situates it among all Data Science Manager salaries.
The salary of USD 54,094 in 2021 is at the bottom not only in 2021 but even over the whole period. The value just above it in 2021 is almost USD 90,000 higher.
Actually, it is very close to a salary equivalent to USD 56,723, reported for a position of Lead Data Scientist at senior level in India in 2022, as shown in this table.
In summary, about employee's residence, it has been seen that
- in the dataset under review, there are seven Data Science positions in the United States with primary countries of residence different from the United States;
- the countries are Brazil, Greece, India, the Philippines, and Russia;
- three positions are in small-size companies, two in medium-size companies, and two in large companies;
- the corresponding salaries are at the very bottom of the corresponding sets of points in six cases, and halfway between the lower fence and the first quartile in one case;
- the three salaries with India as primary country of residence are closer to salaries reported in India than to American medians; the salary with Brazil as primary country of residence is between the American median and estimates for Brazil from the dataset under review or from the Internet.
In a Snapshot: Data Science Salary in the United States
- The United States comes first in salary median and average in the group of Data Science job titles — with a median of USD 150,000 and an average of USD 156,794.
- In 2022, median salary increases from junior to executive level — with most positions being reported at senior level; median at senior level is USD 155,000.
- The median salary for the sole title of Data Scientist at senior level is USD 152,500 in 2022.
- Data Science Managers have a median salary of USD 185,400 in 2022.
- The large company median ranks a little higher than the medium-size company median in 2022; the small-size company median is noticeably lower, with just four positions reported.
- Median salary stands approximately at the same quantitative niveau in the subgroup with remote ratios lower than 20% and in the subgroup with remote ratios greater than 80%; the in-between subgroup scores higher, with just three positions reported.
- In seven cases, the employee's primary residence is in other countries; six out of these cases coincide with salaries at the bottom of the corresponding sets of points in the United States and one is halfway between the lower fence and the first quartile; in four cases, salaries in the primary residence countries are available in the dataset under review and they are indeed lower than in the United States; these factual findings on tiny subsamples need generalizing.
As shown in this inter-country comparison table for Data Science salaries, outside of the United States the number of positions by country is much smaller. Consequently, only a few countries will be analyzed separately.
But Europe offers a broader scope, characterized among others not only by geographical proximity but also by interrelations within the European Union and outside of it. This does not exclude the presence of salary disparities in Europe, as we had also observed inside the United States.
Let us get started with the evolution over time of Data Science salaries in Europe.
The median salary of the group of Data Science job titles follows a caret-shaped evolution from 2020 to 2022, the 2021 median being noticeably higher.
The data distribution are skewed right. The 2021 distribution is rather elongated with a maximum kde of 0.632. The 2022 distribution has no low outlier, only two high outliers.
Some summary statistics about 2022 have been transcribed into the next table. They deserve commenting — among others from an international perspective.
On the basis of the dataset under review, in 2022 the median salary for Data Science in Europe is hardly more than one third of the median salary in the United States. The European median salary is the equivalent of USD 54,634 in 2022, against USD 150,000 in the United States.
In average salary, the European estimate — equivalent to USD 65,101 — represents 42% of the American one, which is USD 156,794.
The European median and average are based on 61 positions reported, against 371 for the United States.
This comparison may raise questions such as the following ones.
On Reddit — consulted on January 24, 2023 — we can read: "Why does the US pay so much more for Data Science?"
On Quora — consulted on January 24, 2023 — another question arises: "What are the differences between a Data scientist's salary in US vs other countries?" In a tentative answer, Claude Lambert refers to differences in cost of living.
These questions and this tentative answer – although crucial from other points of view – fall out of the scope of the current data analysis.
In the scope of this study, the gap between American and European medians for full-time jobs in Data Science sheds some quantitative support to company location being an impactful factor of salary. Such support had already been provided
Let us now turn to the salary breakdown by experience level in Europe.
Median salary by experience level evolves in caret-shaped profile
- Increase is clear from junior level to mid-level.
- It is quite largely canceled when moving from mid-level to senior level, which is the opposite of what is observed for the United States.
- There is no executive level position, which is once again different from the American situation.
- The most populated subsample is the mid-level one — contrary to the situation in the United States, where senior level predominates.
Why is senior level median lower than mid-level median?
- Let us drill down into data by hovering over the sets of points and let us pay attention — among others — to country composition.
- At mid-level, company locations are primarily in the United Kingdom, and secondarily in the Netherlands, Germany, and France.
- At senior level, the violin plot visually stratifies the set of points into three tiers, which correspond not only to salary niveaux but also to different country compositions. At the top, there is one salary relating to Germany. At the second quantitative niveau, there are two positions in the United Kingdom. At the third niveau, a vast majority of positions are registered in Spain – 14 positions easily discernible when using the range slider —, together with one position in Portugal and one position in France.
- The lowest subgroup pulls median salary down in spite of the British and German medians — and averages — increasing from mid-level to senior level. British median and average increasing from mid-level to senior level shows in this table. The increase for Germany shows in this table.
After perusing cross-effects of experience level and country composition in 2022, let us investigate the very job title of Data Scientist on its own; the other Data Science job titles will be tackled afterwards.
The salaries corresponding to the job title of Data Scientist only are split by experience level in the table hereunder.
Just as for the whole group of Data Science job titles, for the job title of Data Scientist on its own, median salary is higher at mid-level.
What is the evolution over time of the median of the most populated subgroup — that is to say the mid-subgroup?
The Data Science job titles other than Data Scientist are detailed in the next table.
The table provides summary statistics by job title and by experience level, with occurrence numbers. Rows are sorted by occurrence number in decreasing order and then by median salary also in decreasing order.
Among the salaries of Data Science job titles other than Data Scientist, in 2022 the highest salary is equivalent to USD 155,499 for a Principal Data Scientist at senior level — in Germany as can be seen by hovering over the top of the senior level set of points in this violin plot at the top of the 2022 point set.
In the interactive table above, there is also a job title of Director of Data Science — in France as can be seen in this violin plot, which is the fourth position in the 2022 point set when scrolling down in the point set.
Since it is a title of director, executive level might be expected on the basis of the official definition of the EX level in the dataset under review, which reads for EX: "Executive-level / Director".
But the position is reported at senior level. This raises terminology questions. Is it the exact same terminology in Europe — in this case in France — as in the United States? Might the total absence of executive level in Europe in 2022 for Data Science be — even marginally — a matter of terminology?
Leaving these questions open, let us switch to the analysis of another candidate factor for explaining salary fluctuations: company size.
The progression of the median salary by company size does not provide any evidence of company size being an influential salary factor in Europe in 2022.
Let us have a look at the corresponding sets of points.
Just as in the United States — at least in the dataset under review — medium-size companies predominate in Europe in 2022. In 2021, there were more positions in the large companies subgroup than in the other two subgroups.
When hovering over the set of points of medium-size companies, we can state that the United Kingdom and Spain are predominantly present in that subgroup and less in the other two. This means that they counter-balance each other in terms of salary level and this contributes the flat profile of median salaries ranked by company size, which is observed in this box plot over breakdown by company size.
On the contrary, in the breakdown by experience level, hovering over this graph and more particularly over the senior level point set shows that the two countries are predominantly present in different subgroups, and that contributes differences in medians — as explicated in this comment.
Let us switch to the breakdown by remote ratio in Europe in 2022.
The median salary of the whole group of Data Science job titles is expressed by remote ratio for 2022 in the next box plot.
The subgroup with remote ratios greater than 80% has a higher median salary. The other two subgroups are practically at the same quantitative niveau.
How are positions distributed among the three subgroups?
All three subgroups are rather evenly populated in 2022, contrary to the situation in the United States where positions were scarce in the intermediate subgroup.
Hovering over the sets of points reveals that the United Kingdom and Spain are mainly in the subgroup having remote ratios smaller than 20%.
In the United States, positions whose primary country of residence differs from the United States coincide with salaries at the bottom of the corresponding sets of points or in one case halfway the lower fence and the first quartile.
The exact same approach is applied to Europe. The next table is comprised of all positions with a difference between the — European — country of company location and the employee's primary country of residence.
There are seven cases.
The salary equivalent to USD 25,532 attracts attention with its quantitatively low niveau. It corresponds to a full-time Data Scientist job at mid-level exercised in 2021 for a company located in Germany by an employee whose primary country of residence is Serbia.
For the sake of comparison, let us have a look at Data Science salaries at mid-level in Germany in 2021.
In the table above, the salary equivalent to USD 25,532 appears at the very bottom of the subset of the salaries earned in 2021 by Data Scientists at mid-level. In this small subset of four salaries, the median salary is the equivalent of USD 89,694 — with three positions close to that quantitative niveau — and the average is the equivalent of USD 73,914.
In a Snapshot: Data Science Salary in Europe
- In the data set under review, the European Data Science median salary — of the equivalent of USD 54,634 — is in a ratio of 1 to 2.7 with respect to the United States in 2022.
- Just as observed inside the United States, there are salary disparities among countries on the criterion of company location.
- As far as breakdown by experience is concerned, median salary follows a caret-shaped evolution; median salary is higher at mid-level than at senior level, notably because jobs in the United Kingdom – with a higher median salary – are concentrated at mid-level, while jobs in Spain – with a lower median salary – are concentrated at senior level.
- Most jobs are reported for medium-size companies in 2022.
- Median salary is somewhat higher with remote ratios of more than 80%.
- In seven cases, the employee's primary country of residence differs from the company location country. In one case, the information available in the dataset under review clearly shows that the salary is closer to the salaries in the employee's primary country of residence than in the company location country.
Here are summary statistics centralized around the median for the salaries of Data Science professional occupations in the United Kingdom in 2022.
In 2022, the median salary for Data Science professional occupations lies at the equivalent of USD 75,111 in the United Kingdom.
In the United States, the corresponding median salary for Data Science jobs is twice as high, namely USD 150,000.
The United Kingdom median lies itself 37% higher than the European median salary, which is equivalent to USD 54,634.
Once again, company location seems impactful.
Let us turn to the salary breakdowns by job title and by experience level for Data Science jobs in the United Kingdom in 2022.
The next graph shows the distribution of Data Science salaries by experience level. Just as in other graphs above, each point is matched in hover information with nine features.
Actually, all positions are reported at mid-level with the exception of three of them. Junior and senior levels are hardly populated. There is no job position at executive level.
Median and average increase with experience but at junior or senior level they are based on just one or two positions...
More insights come from the following table, which is comprised of a cross breakdown between job title and experience level.
In the dataset under review, mid-level Data Scientists predominate in the United Kingdom in 2022, contrasting with the prevalence of senior level Data Scientists in the United States as shown in this graph.
What about frequency by company size?
Just as in the breakdown by experience level, there is concentration in the middle category, that is to say in the medium-size company category – which, by the way, was populated neither in 2021 nor in 2020.
Let us switch to the breakdown by remote ratio.
Job positions are mostly reported with remote ratios of less than 20% for 2022. Not surprisingly, it was the opposite in 2020.
In a Snapshot: Data Science Salary in the United Kingdom
- The British median salary — equivalent to USD 75,111 — is half that in the United States, but 37% higher than the European median.
- Median salary is growing with experience level from junior to senior level, with junior and senior levels hardly populated.
- Most jobs are in medium-size companies.
- Remote ratios below 20% are predominant in 2022, contrary to 2020 and 2021...
While median salary in the United Kingdom is 37% higher than the European median, in Spain median salary is 31% lower than the European median of USD 54,634.
Which is the salary breakdown by job title and by experience level?
Senior level predominates in occurrence, contrary to the situation in the United Kingdom, where mid-level predominates. At senior level, points amass in two subsets, at two quantitative niveaux. The range slider provides readability into the two subsets of points.
Mid-level is empty and at junior level there are only two positions reported.
Median salary grows from junior level — based on two positions — to senior level.
Aggregation by job title and experience level is provided below.
The median salary at senior level is equivalent to USD 42,552 in Spain in 2022. It is far lower than the British median salary equivalent to USD 83,114 for the job title of Data Scientist at the experience level that predominates in occurrence in the United Kingdom — namely mid-level.
The breakdown by company size is provided below.
Almost all jobs are reported for medium-size companies, just as in the United Kingdom.
Last but not least, the breakdown by remote ratio is presented below.
Almost all job positions are reported with remote ratios of less than 20%, which is rather similar to the situation in the United Kingdom.
In a Snapshot: Data Science Salary in Spain
- In 2022, median salary — equivalent to USD 37,824 — is substantially lower than the European median and especially the median of the United Kingdom.
- Almost all positions are reported at senior level, in medium-size companies, and with a remote ratio lower than 20%.
The dataset under review shows the French median salary very close to the European one.
The next table combines job titles and experience levels.
In median salary, the Netherlands is higher than the United Kingdom and all European Union countries. Switzerland is higher, with one position reported.
The four positions reported in the Netherlands are at mid-level, as shown in the next table.
India has already been referred to among others here, in relation with employees having India as their primary country of residence and working for companies located in the United States.
Here are the positions at companies located in India.
This section covers all job titles referring explicitly to Machine Learning or Deep Learning. The job title group will be referred to as "Machine Learning" or "ML" — unless otherwise stated.
Most positions are in companies located in the United States.
Sorting the interactive table above by median salary in descending order shows first Mexico, second Puerto Rico – each of them with 2 positions reported –, third the United States – with 120 positions reported –, then Canada, Switzerland, France, the United Arab Emirates, India, etc.
Sorting the interactive table above by median salary in ascending order shows first Morocco, second Argentina, third Egypt, etc.
The violin graph above shows — among others — the following patterns:
The bimodal distribution — with peaks at different heights — appears clearly in the violin graph above. There are two concentration areas — almost two clusters:
Drilling down into individual data can be easily performed by actioning the range slider and hovering over the set of points. This shows — about 2022 — that
Some summary statistics are transcribed just below into a table, with average and position count.
The average salary — USD 163,885 — for the whole group of Machine Learning jobs titles in the United States in 2022 is close to the median salary — USD 164,996.
About the proximity between median and average, let us remember that no salaries are considered outliers in the set of points of 2022, as shown in this violin plot; there is no lower fence and no upper fence to cordon off some extreme values...
Median salary increases with experience level, but at a changing pace. Indeed, the four experience levels are regrouped at two quantitative niveaux:
Just as in Data Science, the most populated subgroup in 2022 is at senior level.
The lowest two salaries are at mid-level in 2022. Hovering delivers additional information: both salaries are USD 75,000 for Applied Machine Learning Scientists working for respectively a medium-size company and a large company with a remote ratio of more than 80%, and having Bolivia as primary country of residence.
In the Data Science section, differences between company location country and primary country of residence have been listed; comparisons have been tentatively sketched with the salaries prevailing on one side or the other, with a view to disentangling somewhat the possible respective influences. A similar table will be made available below, with the cases in Machine Learning.
Let us get started with the most frequent role in Machine Learning — in the dataset under review —, that is to say the job title of Machine Learning Engineer. The job title of ML Engineer has been regrouped with the job title of Machine Learning Engineer. Median, first and third quartiles, lower and upper fences, minimum and maximum, and average are — hardly or even not at all — affected by regrouping the two job titles.
The violin plot above calls — among others — for the following observations about 2022:
The two peaks of the salary distribution for 2022 are approximately:
For the job titles of Machine Learning Engineer and ML Engineer, some summary statistics have been transcribed into the next table, with average salary and position count.
The average salary of Machine Learning Engineers — USD 168,521 — is lower than the median salary — USD 176,100.
After going through the salaries linked to the very job title of Machine Learning Engineer — and ML Engineer —, let us branch out into the other Machine Learning job titles in the United States.
Besides the job title of Machine Learning Engineer — and ML Engineer —, the other Machine Learning job titles are presented in the next table for 2022 in the United States.
Salary summary statistics are stratified by experience level. Combinations of job titles and experience levels are sorted by occurrence number in descending order — since subsample size contributes statistical reliability.
The table above shows — among others — the following patterns:
After perusing the breakdowns by job title and experience level, let us analyze the link between Machine Learning salary and company size.
The violin plot above allows the following observations:
The relatively substantial density of the medium-size companies subsample contributes the statistical reliability of the corresponding median.
Let us now turn to the breakdown by remote ratio for the whole group of Machine Learning job titles.
The violin plot above shows that in 2022
In 2021 and 2020, there was no position with a remote ratio smaller than 20%.
Here are the cases where the employee's primary country of residence differs from the company location country — that is to say from the United States.
Let us situate these three job positions among all positions of Applied Machine Learning Scientist at mid-level in the United States over the whole period 2020-2022.
Actually, the dataset under review is only comprised of five positions of Applied Machine Learning Scientist at mid-level in the United States, including the three positions already pinpointed with primary residence abroad.
The salaries corresponding to these three positions are at the bottom. Indeed, the two salaries of USD 75,000 in 2022 and the salary of USD 38,400 are far below the other two salaries of USD 423,000 and USD 173,000.
These three salaries are also at the bottom when considering all Machine Learning jobs:
In brief, the three positions with primary residence abroad coincide with salaries at the bottom of the subcategories they are in. These factual findings are based on a tiny subsample and need generalizing.
Just as in Data Science, in Machine Learning the European median salary — equivalent to USD 69,344 in 2022 — is noticeably lower than the American one — USD 164,996 in 2022.
Some summary statistics are transcribed into the next table, with average and position count.
In Machine Learning in Europe, median salary and average salary are relatively close to each other in 2022.
The sample size is rather limited.
As in Data Science, mid-level has the highest median salary, and senior level is substantially lower. The data analysis conducted in Data Science in terms of composition by country of the last two subgroups could be reconducted here, mutatis mutandis.
Salaries are distributed more or less evenly between the three sub-groups.
Here are some aggregate statistics by combination of job title and experience level.
The next table gives, for all combinations of job title and experience level in Machine Learning, the corresponding medians, averages, minima, maxima, and position count.
Among all job positions in Machine Learning with company location in Europe, the next table lists the positions where the employee's primary country of residence differs from the company location country.
Among all job positions in companies located in Europe, in five cases the employee's primary country of residence differs from the company location country.
In these five cases, five countries are mentioned, four job titles, and three experience levels.
Given the diversity of the cases, the tiny size of the subsamples, and the much lower availability of information in the dataset under review and of estimates on the Internet than for wages in the United States, the analysis is not taken any further.
Contrary to what is observed in Data Science, in Machine Learning the Canadian median — equivalent to USD 140,000 — is not dramatically lower than the American median — namely USD 164,996 —, at least on the basis of the dataset under review.
The next table adds average and position number.
Computed on the basis of the dataset under review, the difference between the American and the Canadian averages is much smaller than in Data Science. In Machine Learning, the median salary is USD 163,885 in the United States in 2022 and the equivalent of USD 149,033 in Canada.
Here comes for Canada an exhaustive list of all combinations of job titles and experience levels, with summary statistics.
In Machine Learning, the Mexican median ranks first in the ranking by country in 2022 — based on the dataset under review.
There are two full-time positions of Machine Learning Engineer, at senior level, in a medium-size company located in Mexico, with primary residence in Mexico and remote ratio of more than 80%.
In Machine Learning, the Puerto Rican median arrives second after Mexico and slightly in front of the United States — based on the dataset under review.
There are two job positions.
In both cases, it is a full-time job of Machine Learning Engineer, at senior level, in a medium-size company located in Puerto Rico, with primary residence in Puerto Rico and remote ratio of more than 80%.
In this group are all job titles containing one of the following expressions:
Of course, the job title "Head of Data Science" has been excluded.
There is a rather substantial subsample for the United States.
Sorting the interactive table above by median salary in descending order shows Puerto Rico at the top — based on two positions — just as for Machine Learning. Then come the United States, Canada, Lithuania, the United Kingdom, Greece, etc.
Sorting the interactive table above by median salary in ascending order shows first Slovakia, then Estonia, India, etc.
The violin graph above shows no big difference in median salary between 2021 and 2022 — namely USD 151,500 in 2021 and USD 145,000 in 2022. Median salary was noticeably lower in 2020 — based on a few positions.
About 2022, hovering over the corresponding violin figure shows that
Let us have a look at the small group of top outliers in 2022 by hovering over the top of the 2022 point set and dragging the range slider:
At the bottom of the 2022 point set, there are three salaries a little isolated in a range of USD 24,000 - USD 25,000. The next salaries above are USD 45,600, USD 50,000, ...
Hovering over the sets of points of the three years shows that the job title of Data Engineer predominates every year. Among other job titles, the Data Architect role appears several times at the top of the sets of points in 2021 and 2022. Both job titles will be studied separately.
The next table adds average and position count for 2022.
Median and average salaries are of the same order of magnitude as the ones for Machine Learning and Data Science, somewhat lower though. We notice a relatively low minimum.
The positions reported reach the number of 471.
The most common job title being Data Engineer, let us get started with analyzing salaries corresponding to that role.
The average (global gross) salary of USD 151,115 is in line with Built In estimate — consulted on January 7, 2023 — which is an average total compensation of USD 150,581.
Fifty percent of all Data Engineer salaries are between USD 117,000 and USD 180,000 — the first and third quartiles.
The next table splits Data Engineer salaries by experience level.
The graph above provides numerous pieces of information — and among others the following ones:
At senior level — the most populated level by far as shown in the table just below —
The median salaries by experience level are transcribed into the next table, with position count.
Executive level is predominant in median salary, senior level in position count.
Let us visualize more salary summary statistics about the largest subgroup, namely the subgroup of Data Engineers at senior level.
At senior level, median salary is USD 150,000 — against USD 145,000 for the global median salary of Data Engineers. In fact, the subgroup at senior level constitute by far the largest part of the whole group and is therefore impactful in terms of medians, all the more so because medians are not influenced by extreme values...
Fifty percent of salaries are distributed between USD 120,000 and USD 182,375 — between the first and third quartiles.
The next table allows to take a quick look at the previous year.
In terms of median salary of Data Engineers at senior level, evolution looks like a plateau from 2021 — when four positions were collected — to 2022.
Senior level is the only level represented among Data Architects in the dataset under review.
The median salary of Data Architects at senior level is slightly higher than the median salary of Data Engineers at senior level — namely USD 154,520 against USD 150,000.
All salaries
The salary distribution for Data Architects at senior level is slightly bimodal, with some concentration around the two peaks, wich are
Hovering over the set of points — preferably after dragging the range slider — informs us that most positions are exercised in or for medium-size companies, with remote ratios often greater than 80% or smaller than 20%.
The next table gathers summary statistics, adding average and position count.
Between Data Engineers and Data Architects, salary disparity is a little more pronounced in average than in median: the average is USD 167,207 for Data Architects and USD 151,115 for Data Engineers.
Sorting the interactive table above by maximum salary in descending order shows at the top two salaries of Head of Data:
They are followed by a salary of USD 193,000 for a Data Engineering Manager at senior level.
The following two graphs split salaries by company size for the whole group of Data Engineering job titles.
Just as in Data Science, in Data Engineering median salary is slightly higher in large companies than in medium-size companies.
The small-size companies category is almost empty, as this is confirmed by the next graph.
The graph above shows the prevalence of medium-size companies in 2022, unlike in 2021 when large companies predominated. There is just one position in a small-size company.
What about the breakdown by remote ratio?
Median salary is somewhat higher with remote ratios greater than 80% than with remote ratios smaller than 20%. Both extreme subgroups are substantially populated, with the intermediate subgroup being empty.
Distribution was rather similar in Data Science and Machine Learning, with the in-between subgroup being scarcely populated.
In the Data Science section, several positions with primary residence abroad coincided with salaries at the bottom of the corresponding sets of points or, in one case, halfway the lower fence and the first quartile.
In the Data Engineer section, such a finding has not been made, for instance when hovering over the bottom of the sets of points in the violin graph just above.
But, for the sake of consistency, let us also look in a systemic way at all jobs with another country than the United States as primary country of residence.
Actually, there are seven positions with the primary country of residence differing from the company location country. They will be analyzed below in the following order of job titles:
The seven salaries will be situated among other salaries from the same subcategories by using either information from the dataset under review — when available — or external information from the Internet.
There are two positions of Data Engineer at mid-level with the primary country of residence differing from the company location country. One is in 2020, the other one in 2021.
Let us situate them among the other positions of Data Engineer at mid-level in the following violin graph — which provides extensive information in the form of summary statistics or detailed individual hover information.
The first position — with Spain as primary country of residence — corresponds to a salary of USD 130,800 in 2020. It appears at the top of the small set of points of 2020, noticeably higher than the other two salaries at mid-level — which are USD 106,000 and USD 110,000.
Moreover, it is even above the median salaries at mid-level from 2021 and 2022 — which are respectively USD 110,888 and USD 120,000.
Consequently, the USD 130,800 salary — with Spain as primary country of residence — more than compares to other salaries in similar positions with the United States as the primary country of residence.
On the contrary, the salary of USD 26,005 in 2021 for a mid-level Data Engineer — with Romania as primary country of residence and a remote ratio smaller than 20% — is at the bottom of the set of points of 2021.
There is also a Cloud Data Engineer position in 2021 at senior level with primary residence in Brazil. The salary is USD 160,000.
This USD 160,000 salary bears comparison with Talent 2023 estimates for Cloud Engineers — consulted on January 25, 2023: Talent median is USD 131,418, with entry levels starting at USD 110,022 and most experienced workers reaching up to USD 169,335.
There are two positions of Lead Data Engineer:
Let us compare them with Talent estimates for Lead Data Engineers — consulted on January 11, 2023: Talent median is USD 150,000, with entry levels starting at USD 134,384 and most experienced workers reaching up to USD 185,056.
The salary of USD 56,000 in 2020 is well below the minimum of USD 134,384 displayed by Talent. Would it have been influenced by lower salaries in Portugal in Data Engineering — as shown in this table for 2022?
On the contrary, the salary of USD 160,000 in 2021 — with Puerto Rico as primary residence — seems to bear comparison with Talent estimates for the United States published in 2023. Would it have been influenced by American salaries, or by salaries in Puerto Rico, which stand in average at the top of this inter-country comparison table in Data Engineering on the basis of two positions reported for 2022 in the dataset under review?
There is also a salary of USD 193,000 in 2022 — with Spain as primary country of residence — for a Data Engineering Manager at senior level. Let us situate it among other positions of Data Engineering Manager in the dataset under review.
The graphic above clarifies the debate: the salary of USD 193,000 in 2022 — with Spain as primary country of residence — lies substantially higher than similar positions in 2021 with the United States as primary country of residence.
Generally speaking, the salary of USD 193,000 looks closer to American Data Engineering salaries than to Spanish Data Engineering salaries, on the basis of this inter-country comparison table.
Last, there is a salary of USD 200,000 — with Malaysia as primary country of residence — for a Head of Data at senior level.
Let us apply our general methodology and situate this salary of USD 200,000 among other positions of Head of Data.
The salary of USD 200,000 for a Head of Data at senior level — with Malaysia as primary country of residence — is is the middle of three salaries of Head of Data at executive level — which has often ranked at a quantitatively higher niveau than other experience levels in the United States in the dataset under review...
In Summary about Possible Primary Residence Impact
- There are seven salaries corresponding to foreign countries as primary country of residence.
- Five salaries — with Brazil, Malaysia, Puerto Rico, or Spain as primary country of residence — are in line or above similar salaries with the United States as primary country of residence; in the case of Puerto Rico, it should be mentioned that anyway — at least in the dataset under review — Data Engineering salaries are a little higher in Puerto Rico — based on two observations — than in the United States.
- Two salaries — with Portugal or Romania as primary country of residence — are noticeably below American standards deducted from the dataset under review; in the case of Portugal, the salary is close to salaries in Portugal reported in the dataset under review.
- These are factual findings that would need generalizing.
The violin figures show a v-shaped profile as evolution over time of the median salary in Europe in Data Engineering.
Hovering over the violin figures reveals as median salary the equivalent of
The 2022 median salary equivalent to US 73,546 is barely more than half the American median salary of USD 145,000.
The three data distributions have each their own shape.
Ranking countries by median, average, minimum, or maximum salary can be done in this inter-country comparison table.
If we sort that inter-country comparison table by median salary in descending order, at the top are Lithuania, the United Kingdom, Greece, and Spain — unlike in Data Science.
When sorting countries by median salary in ascending order, first we find Slovakia, and then Estonia, Portugal, and Germany.
The next table transcribes some salary summary statistics about Europe, adding average and position count.
The European average is close to the median salary, on 69 positions reported.
Below, the analysis branches out by breaking down salaries by job title, experience level, company size, remote ratio, and primary country of residence.
Let us get started with the role of Data Engineer — which is by far the most frequent job title in Data Engineering in Europe, in the dataset under review.
With respect to summary statistics for the whole group of Data Engineering job titles, there is very little change when restricting observations to the sole role of Data Engineer — except for the lowest value, which is substantially higher.
A breakdown by experience level follows.
The violin graph above shows — among others — that
Let us take a quick look at job titles other than Data Engineer.
Besides the job title of Data Engineer, there are the job titles of Head of Data and Cloud Data Engineer.
After breaking down salaries by job title, let us opt again for a global approach — with all job titles together — and investigate the link with company size.
As in Data Science, medium-size companies predominate in occurrence number in 2022; there are only three positions in small-size companies and no positions in large companies.
The median salary of medium-size companies is substantially higher than the one of small-size companies, based on three positions.
What about the remote ratio?
Ranked by remote ratio, salaries amass at the extremes, just as in Data Science and Machine Learning. There is just one position in the intermediate subgroup.
Median salary is at the same quantitative niveau in both extreme subgroups.
The last salary feature will be examined now: it is the employee's primary country of residence.
In 95.9% of positions reported, company location and primary residence are in the same country. In these cases, it is hard to discriminate between respective influences. But in a few cases, the company location country and the primary residence country differ; in these cases, are salaries closer to the ones in the company location country or to the ones in the primary residence country?
Let us first list all cases in Data Engineering in Europe.
Out of the seven cases from the table above, the two extreme salaries will be examined:
The next interactive plot situates them vis-à-vis the other positions of Big Data Engineer in Europe.
In 2021, the salary equivalent to USD 114,047 for a Big Data Engineer at senior level with the United Kingdom as company location country and Poland as primary country of residence lies slightly above a salary of Big Data Engineer with the United Kingdom as company location and primary residence.
This is also more than Talent estimates for the salaries of Senior Data Engineers — consulted on February 1, 2023 — which give
By the way, in the graph above, the salary equivalent to USD 114,047 stands much higher than the small set of points of 2022, comprised of salaries of Big Data Engineer at mid-level or junior level in other European countries.
In brief, the salary equivalent to USD 114,047 more than compares to other Big Data Engineer salaries made available here.
At the opposite, the salary equivalent to USD 5,882 in 2021 at junior level with Switzerland as company location country and India as primary country of residence is the lowest of all Data Engineering salaries in Europe over the whole period, as shown in this graph.
This salary might be closer to India's standards, as shown in this graph. But this might need nuancing, for instance if this junior level position was exercised in a scheme for new entrants, such as internship, traineeship, or work-study contract.
Based on the dataset under review, the evolution profile of the median salary looks more like a plateau when comparing 2021 and 2022:
The 2022 median — equivalent to USD 76,958 — is slightly higher than half that of the United States — namely USD 145,000.
Some 2022 summary statistics are transcribed into the next table, with average and position count.
Contrary to what is observed in Data Science, in Data Engineering the British median salary is hardly higher than the European one.
Nevertheless, this comparison needs clarifying. Indeed, the United Kingdom is more prevalent in Data Engineering than in Data Science. Consequently, in Data Engineering, the United Kingdom represents a greater proportion of positions, which can impact the European median and average more than in Data Science.
The breakdown by experience level is shown next for the salaries of the Data Engineering job title group.
The median salary by experience level shows an inverted-v-shaped profile, median salary being substantially lower at junior and senior level than at mid-level.
Once again mid-level predominates in occurrence in the United Kingdom, just as in Europe as a whole and just as in Data Science and Machine Learning. In the United States — let us remember — the predominant subgroup is the senior level one, irrespective of the domain.
In the dataset under review, all Data Engineering jobs in the United Kingdom are exercised in or for medium-size companies in 2022. This predominance also exists in Data Science but in a less marked way.
Median salary is noticeably lower with remote ratios greater than 80% than with remote ratios smaller than 20%.
There is no position with a remote ratio in between 20% and 80%.
As this inter-country comparison indicates, Greece ranks sixth in median salary in Data Engineering in 2022, with a median salary equivalent to USD 73,546.
All job positions in 2022 correspond to the job title of Data Engineer at mid-level.
The 2022 median salary — equivalent to USD 73,546 — is at the same quantitative niveau as for all of Europe.
As previously shown in this inter-country comparison, Puerto Rico is at the top in median salary in Data Engineering as in Machine Learning, with two positions each time.
Puerto Rico's median salary is USD 167,500 in Data Engineering.
This section deals with all job titles containing the expression "Data Analyst".
First glance reveals that, for the United States, central statistics — here median and average — are significantly lower than in the other three domains — Data Science, Machine Learning, and Data Engineering. The American median salary is USD 110,925 in 2022 in Data Analysis.
This phenomenon is also observed — proportionately even more — for India — whose median salary is equivalent to USD 11,178 in 2022, just above Brazil — whose median salary is equivalent to USD 7,034.
For some other countries, the comparison with the other domains needs nuancing:
When sorting the interactive table above by median salary in descending order, Canada comes second, after the United States.
Median salary increases from USD 91,000 for 2021 to USD 110,925 for 2022. In 2022, median salary is substantially lower than in Machine Learning, Data Science, or Data Engineering.
The 2022 data distribution and set of points show that
The next table adds average and position count for 2022.
Average salary is also much lower than in the three previous groups of job titles — namely Data Engineering, Machine Learning, and Data Science.
With a view to differentiating job titles, let us first focus on the most frequent role — namely the job title of Data Analyst —, before branching out to the other roles.
The median salary and average salary of Data Analysts are really close to the median and average of the whole group of Data Analysis job titles.
Here is a breakdown by experience level.
From junior level to executive level, median salary increases but at a changing pace: after a jump from junior level to mid-level, progression looks almost like a plateau.
Just as in the three previous domains — Data Science, Machine Learning, and Data Engineering — senior level predominates in occurrence in the United States.
Which are the other Data Analysis job titles in the dataset under review?
The table above is comprised of four job titles:
There is only one position for each combination of job title and experience level.
The highest salary — USD 150,000 — relates to a position of Business Data Analyst at mid-level.
After perusing the various combinations of job titles and experience levels, let us return to a holistic approach encompassing all job titles.
Just as in the other three domains — Data Science, Machine Learning, and Data Engineering —, in Data Analysis the category of medium-size companies predominates in position count.
But contrary to the other three domains, in Data Analysis the median of the subgroup of large companies is lower than the median of the subgroup of medium-size companies.
Why? By hovering over the set of points and data mining into the hover information, we can see that
Median salary is almost at the same quantitative niveau at the extremes.
Once again, just as in Data Science, Machine Learning, and Data Engineering, positions amass at the extremes in terms of remote ratio.
Let us list all cases of primary residence abroad.
So, in the table above, all experience levels are represented as well as two job titles — Data Analyst and BI Data Analyst.
The four positions will be analyzed in the following order:
Each position will be situated in an appropriate subset of comparable positions.
First, the salary of USD 80,000 in 2021 for a Data Analyst at senior level can be visualized in the next interactive graph.
The salary of USD 80,000 can be situated with respect to the 2021 point set or to the 2022 point set:
The salary of USD 20,000 for a Data Analyst at junior level appears in the next graph.
The salary of USD 20,000 looks like an outlier, alone at the bottom of the set of points from 2022.
Would that position at junior level possibly be part of a special scheme for new entrants, such as internship, traineeship, or work-study contract?
The salary of USD 150,000 for a BI Data Analyst at executive level is situated in the next graph among all BI Data Analyst positions from the dataset under review.
It is indeed the only BI Data Analyst position at executive level and so it needs comparing with positions at other experience levels — at least in the dataset under review.
The salary of USD 150,000 at executive level in 2021 is above the other salaries, which are at middle or junior level...
The salary of USD 36,259 for a BI Data Analyst at mid-level in 2021 shows in the next plot.
The salary of USD 36,259 is significantly lower than the other two.
In brief, out of four positons with primary residence abroad
- two salaries seem significantly lower than comparable salaries with primary residence in the United States,
- one salary is approximately halfway the minimum and the first quartile,
- and one salary is significantly higher than salaries at lower levels of experience;
- of course, once again, it should be added that these factual findings are based on tiny subsamples and need generalizing.
Median salary is substantially lower in Data Analysis than in the three previous domains — Data Science, Machine Learning, and Data Engineering.
More summary statistics are available in the next table for 2022.
In the dataset under review, the number of positions reported in Data Analysis is rather limited with respect to the other domains.
Median is noticeably lower than average, which has been boosted by outliers — as shown in the next graph.
The average is inflated by three outliers at the top — unlike the median, which is insensitive to them. Outliers have been defined here.
The three outliers above the upper fence of USD 92,350 are
There is no lower fence. The lowest value in 2022 is the equivalent of USD 20,000 for a Data Analyst at mid-level in Greece.
Let us now focus on the very job title of Data Analyst — which is by far the most frequent in the domain of Data Analysis in the dataset under review.
Here are some summary statistics.
Salaries are broken down by experience level in the next graph.
Medians increases from junior level — which appears only once — to senior level.
Which are the other job titles and their respective salaries?
As usual, let us analyze the possible impact of the primary country of residence.
There is only one instance of the employee's primary country of residence differing from the company location country.
Is this amount — equivalent to USD 88,654 — closer to salaries in the company location country — namely Denmark — or to salaries in the employee's primary country of residence — namely Greece?
About salaries of Data Analysts in Denmark, let us mention two sources:
About salaries of Data Analysts in Greece, if we refer to our inter_country comparison table, median salary in Greece is the equivalent of USD 31,520 and average salary is the equivalent of USD 31,182. The three salaries correspond to positions of Data Analysts at mid-level.
In brief, the salary equivalent to USD 88,654 seems in line — based on the information consulted — with salaries in Denmark, that is to say in the company location country.
In the dataset under review, Canada has in 2022 a median salary of approximately the equivalent of USD 72,000, against approximately USD 111,000 in the United States.
The disparity between the two countries had almost disappeared in Data Engineering after peaking out in Data Science with a ratio of 1 to 2.1 between the Canadian median and the American one.
In Canada, the median salary of Data Analysts is at the same quantitative niveau as the median salary of the whole group of Data Analysis job titles.
The other job titles are listed in the table below, with median, average, minimum, maximum, and position count.
The rest of the section deals with all Data Analysis job titles regrouped.
Here is the breakdown by experience level.
The graph above shows an ascending evolution of the median salary by experience level.
Almost all positions are reported in the intermediate category by company size, except for two of them, one in each of the extreme categories.
In Data Analysis, there are six positions reported for India, which, in this inter-country comparison table has the lowest median salary but one.
At the end of this data analysis, two tables present the hundred lowest salaries and the hundred highest salaries, which correspond approximately with 10% of positions.
Both tables are comprised of nine types of information that have already been provided in hover information about individual data points. Numerous extreme values have already been commented upon.
Thanks again to ai-jobs.net for its full dataset of global AI/ML salaries. The dataset is both very clear and readily available without any compulsory counterpart.
I also thank the U.S. Bureau of Labor Statistics for in-depth statistics and website data tools which allow to customize the content of the information extracted and to obtain it in highly workable files.
My thanks also go to Built In, Forbes, Indeed, Payscale, Salary Explorer, and Talent for clear and readily available data. I have appreciated their delivering salary summary statistics and adding position counts.
I thank Reddit and Quora for interesting questions and answers.
The dataset under review is comprised of almost 2,000 positions reported, mainly in the United States, but also in Canada, Europe, Asia, ...
Company Location
- The breakdown by company location shows the sharpest disparities.
- So, in 2022, global median salary is approximately the equivalent of USD 24,000 in Asia, USD 63,000 in Europe, and USD 142,000 in the continent America.
- In Data Analysis, India is approximately at the equivalent of USD 11,000 and the United States at USD 110,000.
- In Data Engineering, India is approximately at USD 37,000 and Puerto Rico at USD 168,000.
- Inside the United States, the median salary of Data Scientists is in a ratio of 1 to 3.8 between the lowest-paying metropolitan statistical area (MSA) and the highest-paying MSA.
Time
- We have seen that median salary can vary from year to year.
- For instance, the difference between Europe and the continent America is more marked in 2022 than in 2021: in 2022, it is the equivalent of USD 63,000 in Europe and USD 142,000 in the continent America; in 2021 it is USD 64,000 in Europe and USD 115,000 in the continent America.
Domain
- In Machine Learning, median salary is USD 165,000 in 2022 in the United States.
- Data Science is at a moderate distance with USD 150,000.
- Data Engineering is at USD 145,000.
- But Data Analysis is at a much greater distance with USD 110,000.
Job Title
- Within a domain, job title can matter.
- So, in the United States in 2022, median salary is USD 168,950 for Machine Learning Scientists and USD 183,000 for Machine Learning Engineers, both at senior level.
- In the United States in 2022, median salary is USD 152,500 for Data Scientists and USD 185,400 for Data Science Managers, both at senior level.
Experience Level
- Within a job title, median salary can substantially fluctuate with experience level.
- So, in the United States in 2022, the median salary of Data Engineers almost doubles from USD 120,000 at mid-level to USD 239,000 at executive level.
Company Size and Remote Ratio
- Median salary can also vary according to company size and remote ratio.
- Generally differences remain moderate with the exception of subgroups scarcely populated.
Employee's Primary Country of Residence
- In 4,1% of positions reported, the employee's primary country of residence differs from the company location country.
- For instance, in the United States in 2022, seven such cases appear in Data Science: salaries are the lowest in their respective subcategories in six cases, and between lower fence and first quartile in the seventh case. When information is available, these salaries are closer to salaries in the primary countries of residence.
- But in Data Engineering, on seven other cases, five are in line with positions with primary residence in the United States, while two are below.
- These are factual findings based on tiny subsamples, which need studying on a greater scale.
Breakdowns and Position Counts
- Senior level predominates in position count in the United States and Spain in 2022; globally in Europe, mid-level predominates.
- The predominant subgroup of companies is always the subgroup of medium-size companies both in the United States and in Europe in 2022 except for the domain of Machine Learning in Europe where the subgroup of medium-size companies and the subgroup of large companies are the same size.
- Remote ratio is often less than 20% or more than 80% and rarely in between in 2022 both in the United States and in Europe.
Combining holistic views and information detailed down to individual level can boost both data analysis and communication of results.
Sets of Points and Hover Information to Pinpoint Cross Effects
- Hovering over sets of points and drilling down into individual hover information can help detect — among numerous patterns — cross-effects.
- For instance in Data Science in Europe, salary breakdown by experience level is impacted in a decisive way by the country composition of the sub-groups as thoroughly explicated here.
- Another example is provided by Data Analysis in the United States in 2022: the median salary in the medium-size company subgroup surpasses the median of the large company subgroup, which is mainly populated with junior level.
Data Distribution Shape and Bimodality
- Visualizing the very shape of an empirical data distribution can prove useful, in some cases even essential.
- For instance, a bimodal density function can reveal the existence of two subgroups, which would not appear as such in summary statistics. Several bimodal distributions have been pinpointed in this Data Analysis project; let us mention three of them:
Distribution Shape & Difference between Extreme Values/Ouliers
- Extreme values can be dealt with by ignoring the first decile — the lowest 10% — and the last decile — the highest 10%, which is both fast and simple.
- A more flexible approach is to consider them outliers if they lie above the upper fence or below the lower fence. Let us take a few examples:
- in this rather dense data distribution for 2022, only one low value is considered a low outlier, which seems to make sense;
- in the same distribution for 2022, according to the upper fence criterion, only the values from the small upper subgroup are considered high outliers;
- in this rather elongated data distribution of the subgroup of large companies, there is no lower fence ... and no low value considered a low outlier.
- Even the more flexible approach of fences can benefit from visualizing the shape of the data distribution and the hover information from extreme — and other — values.
Evolution over Time
- Merging periods enlarges samples and can contribute statistical reliability.
- But building up time series can bring insights if temporal patterns are present and if subsamples are big enough.
Central Statistics
- Average is widespread and, for this and other reasons, useful. But median offers the advantage of great insensitiveness to outliers and skewed data, which can be dealt with separately.
- Combining median, first decile, and last but one decile is clear. Combining median, quartiles, and fences can bring some flexibility.
- When dealing with empirical distributions, central statistics — although instrumental — do not dispense with viewing point sets, density functions, or hover information.
Combining holistic views and detailed info
- provides readers with multi-tiered information and results, as explicated just above;
- empowers readers to develop their own analysis and obtain their own findings.
Data Analysis Needs Sharing
- Data Analysis not only needs multi-tiered information and fine-tuned results, it also needs discussing and sharing.
- Interaction helps as expressed by my motto: "Challenged by you, succeeding with you", which you can see on my GitHub account together with my contact information.
- Do not hesitate to get in touch.